def pseudo_peripheral_node(A): """ Algorithm in Saad """ from pyamg.graph import breadth_first_search n = A.shape[0] valence = np.diff(A.indptr) # select an initial node x, set delta = 0 x = int(np.random.rand() * n) delta = 0 while 1: # do a level-set traversal from x order, level = breadth_first_search(A, x) # select a node y in the last level with min degree maxlevel = level.max() lastnodes = np.where(level == maxlevel)[0] lastnodesvalence = valence[lastnodes] minlastnodesvalence = lastnodesvalence.min() y = np.where(lastnodesvalence == minlastnodesvalence)[0][0] y = lastnodes[y] # if d(x,y)>delta, set, and go to bfs above if level[y] > delta: x = y delta = level[y] else: return x, order, level
def pseudo_peripheral_node(A): """ Algorithm in Saad """ from pyamg.graph import breadth_first_search n = A.shape[0] valence = np.diff(A.indptr) # select an initial node x, set delta = 0 x = int(np.random.rand() * n) delta = 0 while 1: # do a level-set traversal from x order, level = breadth_first_search(A, x) # select a node y in the last level with min degree maxlevel = level.max() lastnodes = np.where(level == maxlevel)[0] lastnodesvalence = valence[lastnodes] minlastnodesvalence = lastnodesvalence.min() y = np.where(lastnodesvalence == minlastnodesvalence)[0][0] y = lastnodes[y] # if d(x,y)>delta, set, and go to bfs above if level[y] > delta: x = y delta = level[y] else: return x, order, level
def pseudo_peripheral_node(A): """Find a pseudo peripheral node. Parameters ---------- A : sparse matrix Sparse matrix Returns ------- x : int Locaiton of the node order : array BFS ordering level : array BFS levels Notes ----- Algorithm in Saad """ from pyamg.graph import breadth_first_search n = A.shape[0] valence = np.diff(A.indptr) # select an initial node x, set delta = 0 x = int(np.random.rand() * n) delta = 0 while True: # do a level-set traversal from x order, level = breadth_first_search(A, x) # select a node y in the last level with min degree maxlevel = level.max() lastnodes = np.where(level == maxlevel)[0] lastnodesvalence = valence[lastnodes] minlastnodesvalence = lastnodesvalence.min() y = np.where(lastnodesvalence == minlastnodesvalence)[0][0] y = lastnodes[y] # if d(x,y)>delta, set, and go to bfs above if level[y] > delta: x = y delta = level[y] else: return x, order, level
def pseudo_peripheral_node(A): """Find a pseudo peripheral node. Parameters ---------- A : sparse matrix Sparse matrix Returns ------- x : int Locaiton of the node order : array BFS ordering level : array BFS levels Notes ----- Algorithm in Saad """ from pyamg.graph import breadth_first_search n = A.shape[0] valence = np.diff(A.indptr) # select an initial node x, set delta = 0 x = int(np.random.rand() * n) delta = 0 while True: # do a level-set traversal from x order, level = breadth_first_search(A, x) # select a node y in the last level with min degree maxlevel = level.max() lastnodes = np.where(level == maxlevel)[0] lastnodesvalence = valence[lastnodes] minlastnodesvalence = lastnodesvalence.min() y = np.where(lastnodesvalence == minlastnodesvalence)[0][0] y = lastnodes[y] # if d(x,y)>delta, set, and go to bfs above if level[y] > delta: x = y delta = level[y] else: return x, order, level